ORDER

Plaintiff seeks attorney fees after prevailing on a motion to remand this action to state court under a federal fee-shifting statute. Because 28 USC § 1447 (c) provides for recovery of such fees, the court must decide whether and what to award. The latter calls upon the court to figure the amount of a “reasonable attorney fee.”

Defendants removed this case to federal court July 23, 2003. Doc # 1. On November 25, 2003, the court remanded the case and granted plaintiffs motion for costs and fees, with the amount of the cost and fee award to be determined after further submissions from the parties. Plaintiff submitted a request for costs and attorney fees in the amount of $81,288.96. PI Memo (Doc # 49). According to the analysis below, the court awards plaintiff costs and attorney fees in the amount of $27,956.36.

I

Plaintiffs request for attorney fees results from its successful motion to remand, in which plaintiff argued that the notice of removal was defective and untimely. See PI Mot (Doc # 15). Plaintiff subsequently filed a letter requesting leave to obtain discovery on the issue of the parties’ citizenship to determine whether a lack of federal subject matter jurisdiction also required remand. Letter, William Huckins to Vaughn Walker, October 21, 2003 (Doc # 36). Defendants opposed plaintiffs request for discovery and proposed, quite reasonably, that the court first consider the simple procedural arguments in plaintiffs filed motion before the parties researched and briefed the more complicated jurisdictional matters. Letter, Michael Kirby to William Huckins, October 10, 2003 (Doc # 56, Exh A) at 2-3. Plaintiff declined to follow defendants’ suggestion and spent a considerable amount of unnecessary time and effort researching and briefing the jurisdictional issue.

The court granted plaintiffs motion to remand on the procedural ground that removal was untimely. Order (Doc #47). The court noted that deciding the jurisdictional issue was unnecessary because the procedural ground sufficed to remand. Id. The court also granted plaintiffs request for costs and attorney fees with the amount to be determined after submission of supplemental memoranda.

Plaintiff submitted a request for costs and attorney fees in the amount of $81,288.96. PI Memo (Doc # 49). Broken down more specifically, plaintiff requests the following for the work of its attorneys on the case: 79.1 hours at $450 per hour (“/hr”) for Nicholas Waranoff, 52.3 hours at $335/hr for William Huckins and 71.5 hours at $240/hr for Kathryn Perko. Plaintiff requests the following for the work of the legal assistants: 20.7 hours at $160/hr for Delia Cuenca and 1.1 hours for Anita Manfreda at $95/hr. Finally, plaintiff requests $7,596.96 in internet-based legal research costs.

II

A

Under 28 USC § 1447 (c), the court may award “just costs and any actual expenses, including attorney fees, incurred as a result of the removal.”

Read narrowly, section 1447(c) could bar many plaintiffs who successfully move for remand from recovering their attorney fees. For example, a plaintiff who is represented on a contingent fee basis does not incur an incremental expense for attorney services in moving for remand.

The Ninth Circuit has rejected such a narrow interpretation of section 1447(c). In Gotro v. R & B Realty Group, 69 F.3d 1485 (9th Cir.1995), the Ninth Circuit held that a plaintiff who is represented on a contingent fee basis may recover attorney fees under section 1447(c). The Ninth Circuit concluded that Congress could have had “two possible reasons” for choosing the language “actual expenses * * * incurred.” First, this language distinguishes an award under section 1447(c) “from a punitive award which was associated with the formerly required bad faith finding.” Id. at 1487 . Second, this language distinguishes an award under section 1447(c) “from the former requirement of a bond upon removal.” Id. at 1488 . Accordingly, the Ninth Circuit upheld the district court’s $13,564.05 award, an amount presumably based on a lodestar calculation.

With respect to the amount of an award under section 1447(c), the Tenth Circuit has concluded that section 1447(c) requires courts to determine whether the requested fee is reasonable. See Huffman v. Saul Holdings Limited Partnership, 262 F.3d 1128 (10th Cir.2001). The Tenth Circuit noted that district courts have long presumed that an attorney fee award under section 1447(c) must be reasonable. Id. at 1134 (collecting cases). Pointing to the section 1447(c) language “incurred as a result of removal,” the court observed that “unreasonably high fees are not ‘incurred’ as a result of removal; rather, excessive fee requests flow from, and accumulate by means of, improper billing practices * * Id. at 1135 . Nothing in section 1447(c) requires courts to award unreasonable, even if actually incurred, fees. Ibid. The Tenth Circuit stated: “Our holding is that the statute’s limit on actual fees to those ‘incurred as a result of removal’ requires the district court to conduct some sort of reasonableness inquiry.” Id. If courts must conduct “some sort of reasonableness inquiry,” a lodestar analysis would appear to be appropriate.

The Seventh Circuit has rejected reasonableness in determining attorney fees under section 1447(c). See Wisconsin v. Hotline Indus., Inc., 236 F.3d 363 (7th Cir.2000). The Seventh Circuit focused on the “actually * * * incurred” language in concluding that a court must award the “actual amount of fees incurred,” whether reasonable or not. Id at 368.

But the Seventh Circuit’s decision in Wisconsin appears to conflict with Ninth Circuit law. First, in Gotro , the Ninth Circuit rejected the type of strict construction given section 1447(c) by Wisconsin . The dissent spells out this holding even more clearly than the majority opinion. 69 F.3d at 1489-90 (O’Scannlain, J). The dissenting judge argued that under section 1447(c) plaintiffs who “actually incur litigation costs should be reimbursed while those who have not incurred any costs should have no such claim.” Id. at 1490 . The majority’s rejection oí this proposition signals plainly that whether the plaintiff has or has not paid the attorney and, if so, how much, does not determine whether the court should make an award under section 1447(c). Second, in the Ninth Circuit, “[ujnder a fee-shifting statute, the court must calculate awards for attorneys’ fees using the lodestar method.” Staton v. Boeing Co., 327 F.3d 938, 965 (9th Cir. 2003) (internal quotation omitted). Wisconsin recognizes that section 1447(c) is a fee-shifting statute. Id. at 366 . This points to a “reasonable fee,” as that term is widely interpreted.

Furthermore, the Ninth Circuit’s approach makes sense. An overemphasis on the “actual expenses” phrase in section 1447(c) would, as Gotro noted, deny to a contingent fee plaintiff recovery for an improper removal. That same overemphasis might also permit a fee-paying (and presumably deep pocket) plaintiff to run the attorney clock and hence impose an undue fee award on a defendant that incorrectly, but with a plausible basis, removed to federal court. Section 1447(c) is best read as calling for the award of reasonable attorney fees incurred as a result of removal.

In the Ninth Circuit, courts use the lodestar method to calculate reasonable attorney fee awards under fee-shifting statutes and the court will do so here.

B

As detailed in the Supreme Court’s recent decision in Gisbrecht v. Barnhart, 535 U.S. 789 , 122 S.Ct. 1817 , 152 L.Ed.2d 996 (2002), courts’ acceptance of the lodestar method dates from the mid-1970s. See id. at 801 (collecting cases). The lodestar method “achieved dominance in the federal courts” after three Supreme Court decisions in the 1980s, Hensley v. Eckerhart, 461 U.S. 424 , 103 S.Ct. 1933 , 76 L.Ed.2d 40 (1983), Blum v. Stenson, 465 U.S. 886 , 104 S.Ct. 1541, 79 L.Ed.2d 891 (1984), and Pennsylvania v. Delaware Valley Citizens’ Council for Clean Air, 478 U.S. 546 , 106 S.Ct. 3088 , 92 L.Ed.2d 439 (1986). See Gisbrecht, 535 U.S. at 801 , 122 S.Ct. 1817 .

Before the lodestar method was developed, the Ninth Circuit applied the twelve-factor test set forth in Kerr v. Screen Extras Guild, Inc., 526 F.2d 67, 70 (9th Cir.1975), to determine attorney fee awards in fee-shifting cases. The lodestar method greatly simplified this analysis by limiting the court’s consideration to two factors: the reasonableness of the number of hours and the reasonableness of the hourly rate. In sifting through the remains of the Kerr factors, courts have determined that at least five of the Kerr factors have been deemed “subsumed in the initial lodestar calculation.” Morales, 96 F.3d at 363-64. These factors include “(1) the novelty and complexity of the issues, (2) the special skill and experience of counsel, (3) the quality of representation, * * * (4) the results obtained, and (5) the contingent nature of the fee agreement.” Id. at 364 n. 9 (internal quotation and citations omitted; alteration in original). Although the subsumed Kerr factors might be helpful in analyzing reasonableness under the lodestar method, the analysis must now focus on what constitutes a reasonable attorney fee in light of case law subsequent to Kerr . See Fischer v. SJB-P.D., Inc., 214 F.3d 1115 , 1119 n. 3 (9th Cir.2000). Indeed, at least one of the “subsumed” Kerr factors, “the contingent nature of the fee agreement,” has been flatly rejected by the Supreme Court. See City of Burlington v. Dague, 505 U.S. 557 , 112 S.Ct. 2638 , 120 L.Ed.2d 449 (1992). Furthermore, other “subsumed” Kerr factors have been narrowed by Supreme Court precedent subsequent to Kerr , as the court explains more fully below.

The lodestar method is a two-step procedure for determining a “reasonable attorney fee.” After the court calculates the “lodestar figure” by “multiplying the number of hours the prevailing party reasonably expended on the litigation by a reasonable hourly rate,” Morales v. City of San Rafael, 96 F.3d 359, 363 (9th Cir.1996), the court then considers whether “it is necessary to adjust the presumptively reasonable lodestar figure on the basis of the Kerr factors that are not already subsumed in the initial lodestar calculation.” Id. Because the lodestar figure is presumptively reasonable, adjustments should be made only in rare cases. Pennsylvania, 478 U.S. at 565 , 106 S.Ct. 3088 . As plaintiff has provided no justification for adjusting the lodestar figure, the court need not reach the second step.

A reasonable attorney fee is the number of hours and the hourly rate that would be billed by “reasonably competent counsel.” Venegas v. Mitchell, 495 U.S. 82, 86 , 110 S.Ct. 1679 , 109 L.Ed.2d 74 (1990); Blanchard v. Bergeron, 489 U.S. 87 , 109 S.Ct. 939 , 103 L.Ed.2d 67 (1989). In Venegas and Blanchard , the reasonable fee awarded by the district court differed from the fee due under the agreement between the fee applicant and the attorney. In each case, the party entered into a contingent fee agreement, prevailed on the merits and obtained an award of reasonable attorney fees. In Blanchard , the court-awarded fees were greater than the amount due under the fee agreement, whereas, in Venegas , the court-awarded fees were less than the amount due under the fee agreement. In each case, the Supreme Court concluded that the fee agreement was enforceable and did not alter the amount awardable as a reasonable attorney fee. Blanchard, 489 U.S. at 96 , 109 S.Ct. 939 (concluding that the “trial judge should not be limited by the contractual fee agreement between plaintiff and counsel”); Venegas, 495 U.S. at 90 , 110 S.Ct. 1679 (holding that “ § 1988 controls what the losing defendant must pay, not what the prevailing party must pay his lawyer”).

Under Venegas and Blanchard , fee applicants are entitled to an award sufficient to “enable them to secure reasonably competent counsel,” but are not entitled to an award “necessary to secure counsel of their choice.” Venegas, 495 U.S. at 89-90 , 110 S.Ct. 1679 . Accordingly, courts award the fee that would be charged by reasonably competent counsel, not the fee due under the particular agreement between the fee applicant and its attorneys. Limiting the award to the fee charged by reasonably competent counsel fulfills the aim of fee-shifting provisions, which is to allow parties to employ reasonably competent counsel “without cost to themselves if they prevail.” 495 U.S. at 86 , 110 S.Ct. 1679 . Thus, even if a party chooses to employ counsel of unusual skill and experience, the court awards only the fee necessary to secure reasonably competent counsel.

Reasonably competent counsel bill a reasonable number of hours. Reasonably competent counsel do not bill hours that are “excessive, redundant, or otherwise unnecessary.” See Hensley, 461 U.S. at 434 , 103 S.Ct. 1933 . Additionally, the court must take into consideration discounts commonly given to clients and an attorney’s ability to collect fees from its client. As emphasized by the Supreme Court in Hensley :

Hensley, 461 U.S. at 434 , 103 S.Ct. 1933 (internal quotation omitted; emphasis omitted).

Reasonably competent counsel bill at a reasonable hourly rate based on the local legal community as a whole. The Supreme Court’s decision in Blum v. Stenson, 465 U.S. 886 , 104 S.Ct. 1541 , 79 L.Ed.2d 891 (1984), is instructive. In Blum, the prevailing parties were represented by attorneys from “The Legal Aid Society of New York.” 465 U.S. at 891 , 104 S.Ct. 1541 . The prevailing parties sought attorney fees at hourly rates equivalent to those charged by private counsel. The defendants argued to the Supreme Court that “market rates reflect the level of compensation necessary to attract profit making attorneys, but that such rates provide excessive fees to nonprofit counsel.” Id at 893, 104 S.Ct. 1541 . The Supreme Court rejected this argument, concluding that a court should not vary its calculation of a reasonable hourly rate whether the “plaintiff was represented by private counsel or by a nonprofit legal services organization.” Id at 894, 104 S.Ct. 1541 . The Supreme Court also discouraged the use of an attorney’s practice area as a basis for distinguishing an attorney fee award. Id at 893, 104 S.Ct. 1541 (requiring that “the amount of fees awarded under [section 1988] be governed by the same standards which prevail in other types of equally complex Federal litigation, such as antitrust cases”) (internal quotation omitted; alteration in original).

Thus, under Blum, courts are discouraged from using either an attorney’s status as a non-profit, government or private attorney or an attorney’s practice area to justify high hourly rates.

Although the cases cited above arise primarily in the context of contingent fee agreements in civil rights litigation, “the definition of what is a reasonable fee applies uniformly to all federal fee-shifting statutes.” Anderson v. Director, Office Workers Compensation Programs, 91 F.3d 1322, 1325 (9th Cir.1996). The lodestar analysis is applied to attorney fee provisions contained in a large number of statutes covering a broad range of subject areas. See, e g, Solid Waste Disposal Act, § 7002(e), 42 USC § 6972 (e); Civil Rights Attorney’s Fees Awards Act of 1976, 42 USC § 1988 ; Lanham Act, 15 USC § 1117 (a); see also Federal Judicial Center, Awarding Attorneys’ Fees and Managing Fee Litigation 1 (1994) (noting that by 1994 nearly 200 civil statutes authorized the award of attorney fees). The court has been provided no evidence that Congress intended “reasonable” fees under one statute to exceed “reasonable” fees under another statute. Indeed, a court-ordered award of attorney fees at an above-average rate would send an improper expressive message in that the court would essentially be finding that certain types of cases are more valuable than others.

In sum, a reasonable attorney fee is the fee that would be charged by reasonably competent counsel, not counsel of unusual skill and experience. Reasonably competent counsel bill a reasonable number of hours at a reasonable hourly rate. A reasonable hourly rate is based on rates charged in the local legal community as a whole, not particular segments of the bar.

C

In the previous subsection, the court demonstrated that a reasonable attorney fee is the fee billed by reasonably competent counsel, not the fee a particular attorney bills a particular client. In almost all cases, including this one, fee applicants provide data about fees charged by a particular attorney, as well as a general statement that such a rate is similar to the rates charged by some subset of attorneys in the area. The court considers how to convert the data typically provided by fee applicants to the data necessary to calculate a reasonable attorney fee.

First, the court must determine whether the requested number of hours is greater than, less than or the same number of hours that reasonably competent counsel would have billed. If the requested number of hours is greater than the number of hours reasonably competent counsel would have billed, then the court should reduce the requested number of hours accordingly. See Hensley, 461 U.S. at 434 , 103 S.Ct. 1933 (describing the court’s duty to eliminate hours that are “excessive, redundant, or otherwise unnecessary”). If the requested number of hours is less than the number of hours reasonably competent counsel would have billed, the court should compensate the fee applicant at an above-average hourly rate. If the requested number of hours is the same as the number of hours reasonably competent counsel would have billed, the court should use the number of hours requested.

Second, the court must determine a reasonable hourly rate. Cases describing the lodestar calculation provide little guidance in determining an appropriate hourly rate. Courts are directed to compare the requested rates with the “prevailing market rate,” which is the rate “prevailing in the community for similar services of lawyers with reasonably comparable skill, experience, and reputation.” Blum, 465 U.S. at 895 n. 11, 104 S.Ct. 1541 .

Most courts applying this standard come to the general conclusion that a particular attorney’s billing rate is relevant, but not dispositive, evidence of a reasonable hourly rate. See, e g, Guam Society of Obstetricians and Gynecologists v. Ada, 100 F.3d 691, 702 (9th Cir.1996); United Steelwork ers of America v. Phelps Dodge Corp., 896 F.2d 403, 407 (9th Cir.1990); Moore v. Jas. H Matthews & Co., 682 F.2d 830, 840 (9th Cir.1982); see also Gay Officers Action League v. Puerto Rico, 247 F.3d 288, 296 (1st Cir.2001) (“[T]he court may take guidance from, but is not bound by, an attorney’s standard billing rate.”); Crescent Publishing Group, Inc. v. Playboy Enterprises, Inc., 246 F.3d 142, 151 (2d Cir.2001); Public Interest Research Group of New Jersey, Inc. v. Windall, 51 F.3d 1179, 1185 (3d Cir.1995); Louisiana Power & Light Co. v. Kellstrom, 50 F.3d 319, 328 (5th Cir.1995); Pinkham v. Camex, Inc., 84 F.3d 292, 294 (8th Cir.1996); Tire Kingdom, Inc. v. Morgan Tire & Auto, Inc., 253 F.3d 1332, 1337 (11th Cir.2001) (per curiam); Covington v. District of Columbia, 57 F.3d 1101 (D.C.Cir.1995); but see Mathur v. Bd. of Trustees of Southern Illinois Univ., 317 F.3d 738, 743 (7th Cir.2003) (holding that “[o]nly if an attorney is unable to provide evidence of her actual billing rates should a district court look to other evidence, including rates similar experienced attorneys in the community charge paying clients for similar work.”).

Case law provides little guidance in determining what constitutes the “prevailing market rate.” Blum discourages use of two of the categories most often cited in support of an above-average attorney fee award. These categories are an attorney’s status as a private, non-profit or government attorney and an attorney’s practice area. For example, fee applicants often attempt to describe the prevailing market rate as the rate charged by attorneys who work for large and prestigious private law firms. This is exactly what plaintiff has done here. See Rosen Decl (Doc # 50). Plaintiff submits the hourly charges claimed by large, well known and highly regarded law firms. Id. The prevailing market rate, however, is based on rates charged in the local legal community as a whole, not particular segments of the bar.

Additionally, the purpose of using the prevailing market standard can be misunderstood. The purpose of using prevailing market rates is to estimate the hourly rate reasonably competent counsel would charge. The purpose is not to determine whether or not a specific attorney could command a specific hourly rate in the market. In performing a lodestar calculation, the court ensures that the fee applicant receives, and the losing party pays, a reasonable attorney fee; the court need not ensure that the agreement between the fee applicant and its attorneys provides a fair market rate for the attorneys’ services. As stated in Venegas , a court’s determination of a reasonable attorney fee “controls what the losing defendant must pay, not what the prevailing party must pay his lawyer.” Venegas, 495 U.S. at 90 , 110 S.Ct. 1679 .

Accordingly, the average market rate in the local legal community as a whole is a better approximation of the hourly rate that would be charged by reasonably competent counsel than the actual billing rate charged by a single attorney. Like the hypothetical “reasonably competent attorney,” attorneys billing at the average rate will not be unusually skilled or experienced but those attorneys typically capable of rendering the required services. Further, an average market rate combines the rates charged by private, non-profit and government attorneys from a variety of practice areas.

While the average market rate normally governs a fee award, an above-average rate may be appropriate in certain situations. A fee applicant should neither be rewarded for hiring expensive legal counsel nor penalized for hiring more efficient legal counsel. Thus, if a fee applicant can demonstrate that its attorneys billed fewer hours than reasonably competent counsel would have billed, the fee applicant should be reimbursed at an above-average hourly rate. See Blum, 465 U.S. at 898 , 104 S.Ct. 1541 . For example, consider a situation in which a court determines that reasonably competent counsel would have billed 10 hours at $100/hr. The fee applicant in such a case would be entitled to $1000 in fees. 'If the fee applicant in this example were to submit a request for 8 hours, the court should use an hourly rate of $125/hr so that the total fee awarded equals $1000, which is the “reasonable attorney fee.”

Yet this adjustment to reflect above-average attorney efficiency does not suggest that an attorney who, as a result of unusual skill or experience, achieves an above-average result for the client also justifies an above-average hourly rate. An attorney who obtains an above-average award for the client might deserve reimbursement at an above-average hourly rate. But, as presently explained, such a premium over the reasonable rate should come from the client, not at the expense of the losing party. An award of reasonable attorney fees “controls what the losing defendant must pay, not what the prevailing party must pay his lawyer.” Venegas, 495 U.S. at 90 , 110 S.Ct. 1679 . The losing party is obligated to pay only the rate charged by reasonably competent counsel. Accordingly, the losing party should not be obligated to pay an above-average attorney fee award simply because the prevailing party obtained an above-average result by employing counsel of unusual skill or experience. Instead, a party that receives an above-average result should be compensated for the work of reasonably competent counsel.

Nor can an above-average hourly rate be warranted on the basis that counsel of unusual skill and experience increase the chance that a party will prevail. The reasoning of the Supreme Court’s decision in City of Burlington v. Dague (“Burling ton”.), 505 U.S. 557 , 112 S.Ct. 2638 , 120 L.Ed.2d 449 (1992), supports this principle. In Burlington, the Supreme Court held that the lodestar figure- may not be multiplied by a “contingency enhancement,” which is a multiplier based on the ex ante probability that a party will not prevail.

Burlington thus supports limiting the reasonable attorney fee award to the hourly rate charged by reasonably competent counsel in primarily two ways. It bears repeating, as Burlington emphasizes, that a reasonable attorney fee award “controls what the losing defendant must pay, not what the prevailing party must pay his- lawyer.” Venegas, 495 U.S. at 90 , 110 S.Ct. 1679 . The respondent argued thát the lodestar figure should be enhanced “to reflect the fact that the party’s attorneys were retained on a contingent fee basis and thus assumed the risk of receiving no payment at all.” Burlington, 505 U.S. at 559 , 112 S.Ct. 2638 (emphasis added). Such an argument is irrelevant because a contingency fee arrangement governs what the prevailing party must pay the attorney; the contingency arrangement does not govern what the court requires the losing party to pay the prevailing party. Accordingly, the court does not adjust the reasonable attorney fee to be paid by the losing side for the risks in an attorney’s litigation portfolio.

Furthermore, “just as the statutory language limiting fees to prevailing (or substantially prevailing) parties bars a prevailing plaintiff from recovering fees relating to claims on which he lost, so should it bar a prevailing plaintiff from recovering for the risk of loss.” Id. at 564, 112 S.Ct. 2638 (emphasis added). Allowing a contingency enhancement would create unintended problems, including “indiscriminately encouraging nonmer- itorious cases” and effectively shifting all of the fee applicant’s attorney fees to the other side whether or not the fee applicant prevails. Id. at 563 , 112 S.Ct. 2638 . The essential problem with awarding a contingency enhancement is that it double counts the ex ante probability of losing.

A numerical example might be helpful. Consider 10 cases. In each of these cases, the plaintiff has an 80% chance of succeeding if $10,000 in attorney fees are incurred. From an ex ante perspective, a plaintiff considers what he expects to pay in attorney fees. In a fee-shifting case, the expected payment is the cost of attorney fees against the expected value of an attorney fee award if the party prevails. Without a contingency enhancement, each plaintiff expects to pay $2,000 in attorney fees ($10,000*(0.8) — $10,000 = $8,000 — $10,000 = -$2,000). A contingency enhancement can be quantified as the inverse of the likelihood of success (1/0.8). With a contingency enhancement, each plaintiff thus expects to pay $0 in attorney fees ($10,-000*(0.8)*(1/0.8) — $10,000 = $10,000— $10,000 = $0).

The ex ante example demonstrates that a contingency enhancement effectively shifts the risk of loss from the plaintiff to the defendant. A risk is the likelihood of an event multiplied by the magnitude of the event. In this example, the ex ante risk of loss is the likelihood of not prevailing (0.2) multiplied by the amount of attorney fees incurred on a losing claim ($10,000). Whether or not a contingency enhancement is used, the plaintiff bears a risk of loss of $2,000 (-$10,000*(0.2)). But if a contingency enhancement is allowed, the risk of loss is offset perfectly by the “risk” of an enhanced attorney fee award ( [$10,000*(l/0.8) — $10,000]*(0.8) = [$12,-500 — $10,000] *(0.8) = $2,000).

This can be seen from the ex post perspective. Assuming that the odds play out, without a contingency enhancement, 8 of 10 plaintiffs would get $10,000 and 2 of 10 would get $0. With a contingency enhancement, 8 of 10 plaintiffs would get $12,500 and 2 of 10 would get $0. As a group, therefore, the plaintiffs recover their total attorney fees ($12,500*8 4- $0*2 = $10,000*10 = $100,000). On the other side, the defendants as a group pay the plaintiffs’ total attorney fees, the losers’ as well as the winners’ fees.

Thus, with a contingency enhancement, the defendants individually expect to pay each plaintiffs total attorney fees from an ex ante perspective and as a class do pay the plaintiffs’ total attorney fees from an ex post perspective. The Supreme Court described this phenomenon as “indiscriminately encouraging nonmeritorious claims.” Burlington, 505 U.S. at 563 , 112 S.Ct. 2638 .

A contingency enhancement impermissi-bly double counts the ex ante probability of losing. An “enhancement” for the ex ante probability of losing should be applied to the ex ante expected attorney fee recovery, not the ex post award, because once a plaintiff has prevailed, he no longer bears the risk of losing. The ex ante expected award is $8,000 ($10,000*(0.8)). An appropriate “enhancement” would require the court to multiply $8,000 by (1/0.8) to obtain $10,000. A contingency enhancement is unwarranted because it counts the ex ante probability of losing twice ($8,000*(l/0.8)*(l/0.8)).

Accordingly, Burlington bars expressly shifting the ex ante risk of loss.through contingency enhancements to the lodestar figure. The reasoning in support of Burlington also suggests that padding the lodestar figure to adjust for the risk of loss is also barred. Although the “difficulty of establishing” the merits of a case might support a “higher number of hours” or a “higher hourly rate,” the court must avoid unjustly shifting the risk of loss to the losing side. See Burlington, 505 U.S. at 562 , 112 S.Ct. 2638 . It makes little difference to the losing side whether the risk of loss is explicitly shifted by a contingency enhancement or implicitly shifted through incorporating a large number of hours or a high hourly rate into the lodestar figure.

Much of what plaintiff has submitted to support its fee application would lead the court to a result the Supreme Court rejected. The experience of plaintiffs counsel, his high regard among peers in the bar and the high fees charged by the elite segments of the bar, to which plaintiffs papers attest, simply do not bear on the amount of a “reasonable fee.”

Rather, these premia reflect what counsel’s clients are willing to pay to secure the services of a particular lawyer and thus diminish the risk of loss or for a variety of subjective reasons. Conversely, a lawyer may agree to take on a matter and charge a below-average rate because of the importance of the case, its notoriety and a host of other intangible factors. These considerations are properly reflected in the agreement between the lawyer and the client, but not in the court’s award of a fee. A “reasonable fee” award does not necessarily equate to what the client agrees to pay.

In sum, to convert the data provided by fee applicants to a “reasonable attorney fee,” the court first compares the requested number of hours to the number of hours that “reasonably competent counsel” would have billed. The court then multiplies the appropriate number of hours by the average market rate, unless an above-average hourly rate is justified by the unusual efficiency of a party’s attorneys.

The court acknowledges that the approach outlined in this order focuses more on the requested number of hours for all the attorneys in the case than the hourly rates of the individual attorneys. Focusing on the requested number of hours as a whole rather than the requested hourly rates of individual attorneys produces better attorney fee decisions. An attorney’s hourly rate may be based on a host of subjective factors, such as her personality or status as a non-profit, government or private attorney, that are not relevant in calculating a “reasonable attorney fee.” The court is in a better position, however, to determine whether the requested number of hours is objectively reasonable because the court has seen the nature of the factual and legal questions in the case, as well as the quality of the attorneys’ work. Furthermore, the court is concerned with the overall result obtained, not the particular allocation of labor that a party’s attorneys may choose.

Ill

A

Applying the methodology developed in the preceding section, the court first considers whether the requested number of hours is greater than, less than or the same number of hours that reasonably competent counsel would have billed.

Plaintiffs attorney fee request is based on its successful motion to remand. Generally, motions to remand are not complex. In this case, the legal and factual analysis supporting the court’s decision to remand was particularly straightforward. See Order (Doc # 47).

Plaintiff makes a number of general arguments to support the large number of hours it requested. First, plaintiff claims that opposing counsel “professed considerable legal experience and stated he was ‘very familiar’ with the law applicable to removal.” PI Memo (Doc # 49) at 3. Plaintiff also points to a letter in which defen dants’ counsel suggested that plaintiff should “keep Rule 11 in mind” because plaintiffs proposed jurisdictional argument was based on a factual claim that contradicted the allegations of the FAC. Letter, Michael Kirby to Nicholas Waranoff, July 25, 2003 (Doc #51, Exh 8). From this, plaintiff concludes that he was required to be “extremely careful and thorough in researching the issues presented in the remand motion and on reply.” PI Memo (Doc # 49) at 3.

Plaintiff is represented by able and experienced attorneys. Neither a claim by opposing counsel of “considerable legal experience” nor an innuendo regarding FRCP 11 would require plaintiffs counsel to expend an unusually large number of hours on the motion.

Second, plaintiff contends that considerable time and effort was required to reply to defendants’ “lengthy opposition (i.e., 24 pages),” which contained “numerous factual and legal issues.” PI Memo (Doc # 49) at 3. But defendants’ response was not unusually long. Under the Civil Local Rules, non-moving parties are allowed 25 double-spaced pages in which to oppose a motion. Civ LR 7-3, 7-4. Nor is the court persuaded that defendants raised legal or factual issues of unusual complexity.

Finally, plaintiff contends that the motion to remand was of unusual complexity because it involved “serious questions about the citizenship of some of the defendants.” PI Memo (Doc # 49) at 3. The court finds this argument unpersuasive on the facts of this case. The procedural question in this case was relatively straightforward. Plaintiff interjected a more complex jurisdictional issue but defendants reasonably suggested that the court first consider the procedural grounds before the parties tackled the more complex jurisdictional issue. See Letter, Michael Kirby to William Huckins, October 10, 2003 (Doc # 56, Exh A). Plaintiff declined to agree with defendants’ reasonable suggestion and incurred significant unnecessary costs in researching and briefing the jurisdictional issue. Because plaintiff proceeded despite defendants’ suggestion, plaintiffs reliance on the complexity of the jurisdictional issue is unpersuasive.

B

Plaintiff groups its requested hours into four categories: (1) time spent researching and writing its motion and reply, (2) time spent in discovery-related activities, (3) other time spent and (4) time spent in making this fee request. Although the court considers plaintiffs requested hours under these four categories, the court notes that the American Bar Association’s “Litigation Code Set” provides a more uniform methodology for categorizing requested hours. See American Bar Association (“ABA”), Uniform Task-Based Management System Information, available at http://www.abanet.org/litigation/utbms/home.html. The ABA template commends itself to parties applying for fee awards.

1

Plaintiff requests 164.1 hours for time spent by three attorneys researching and writing its motion to remand and reply.

Considering first the time spent on plaintiffs motion to remand, plaintiff requests 86.7 hours of attorney time. This represents about 12 days of an attorney billing 7.5 hours per day. It seems to the court that it would take reasonably competent counsel a week of five days and, perhaps to .be generous, a weekend for good measure, to complete the amount of work evident to the court. To be sure, this is a rough and imprecise estimation, but some appraisal is required. Hence, the court concludes that 7 days of 7.5 hours billed per day is a reasonable number of hours. This yields a total of 52.5 hours on the motion to remand.

Plaintiff also requests 77.4 hours for time spent by attorneys replying to defendants’ opposition. This is a full 10 days-plus of attorney time, which is also excessive in light of the straightforward nature of the motion and opposition. Indeed, a large amount of the effort expended by plaintiffs counsel appears to have been directed to the jurisdictional issue. As described above, plaintiffs efforts were unnecessary and, therefore, the court concludes that some of these hours should be excluded from the lodestar figure. It would be quite unjust to force defendants to pay for an expense that would have been eliminated had plaintiff followed the reasonable alternative proposed by defendants. The court concludes that reasonably competent counsel could have completed this task in 5 days at 7.5 hours billed per day. This yields a total of 37.5 hours on the reply.

Accordingly, the court finds a total of 90 hours for research and writing the motion and reply to be reasonable.

2

Plaintiff requests 23.9 hours for discovery efforts to substantiate the jurisdictional argument. A significant portion of these hours occurred before defendants’ counsel suggested that the jurisdictional issue be tabled. But a significant portion of these hours occurred after the suggestion of defendants’ counsel. Accordingly, the court excludes those hours that plaintiffs counsel spent after they had considered and rejected the suggestion of defendants’ counsel. In sum, the court allows 13.7 hours and does not allow 10.2 hours.

3

Plaintiff requests 5.4 hours in “other fees.” All but 1.6 of these hours were spent in preparing an administrative request for the court to seal two tabs in the record submitted with the notice of removal. Admin Req (Doc # 10); Summary (Doc # 51, Exh 7). Plaintiff requested that the court seal two documents that contained harsh remarks in a settlement conference. Although the court acceded to plaintiffs request to seal the documents, the administrative request appeared to be based on a tangential disagreement between the parties and the attorneys. Accordingly, the court finds that the fees for bringing the request to seal are not materially related to the removal and allows only 1.6 hours of the 5.4 hours requested.

4

Plaintiff requests 9.5 hours of attorney time spent in preparing this request for attorney fees. The court finds the amount of time to be reasonable.

5

Finally, plaintiff requests 21.8 hours of time spent by legal assistants on this case. Under analogous fee-shifting statutes, the Supreme Court has held that the fees of legal assistants may be recovered. Missouri v. Jenkins, 491 U.S. 274, 285 , 109 S.Ct. 2463 , 105 L.Ed.2d 229 (1989). Accordingly, the court concludes that fees for legal assistants may be recovered under section 1447(c). The court finds the amount of time requested by plaintiff to be reasonable.

6

In sum, the court finds that the following number of hours may reasonably be included in the lodestar calculation:

1. Time spent researching and writing motion and reply: 90 hrs;

2. Time spent in discovery: 13.7 hrs;

3. Time spent in other activities: 1.6 hrs;

4. Time spent researching and writing the attorney fees request: 9.5 hours and

5. Time spent by legal assistants: 21.8 hours.

Accordingly, the court allows, in total, 114.8 attorney hours and 21.8 legal assistant hours.

IV

A

The court now turns to determining a reasonably hourly rate. In two recent orders the court has calculated an average market rate in the local legal community as a whole using public data from the United States Census Bureau and Bureau of Labor Statistics (“BLS”). See Allen v. BART, 2003 WL 23333580 (ND Cal, Jul 31, 2003); Gilliam v. Sonoma City, 2003 WL 23341211 (N.D.Cal., Dec 22, 2003). The BLS provides data on the hourly wages earned by attorneys and legal assistants. To estimate the hourly rates billed to clients, the court first calculated the ratio of net receipts to gross receipts from data compiled by the Census Bureau. This ratio was used to approximate the overhead costs that would be incorporated in the hourly rates billed to clients. The court then divided the BLS wage data (w) by the ratio of net receipts (nr) to gross receipts (gr) to determine an estimated average market rate (r) as described in the following equation:

r = w = w* gr nr/gr nr

This methodology produced an approximate average market rate of $150/hr in Allen and $190/hr in Gilliam . The court notes that its previous order in this case was based on the now-outdated data from Allen .

The most recent census data describing gross and net receipts by law partnerships is located in “Statistical Abstract of the United States: 2003” (“2003 Statistical Abstract”). See United States Census Bureau, Statistical Abstract of the United States: 2003, tbl 733, available at http://www.census.gov/statab/www/. The 2003 Statistical Abstract provides gross and net receipts for the year 2000. Gross receipts totaled $79 billion and net receipts totaled $29 billion. This yields a ratio of net receipts to gross receipts of 0.367. The figures for previous years are presented below in Table 1.

Table 1. Gross and Net Receipt Data from the 2001, 2002 and 2003 Statistical Abstracts of the United States.

Publication year Data year Gross receipts-Partnerships (billion $/yr) Net receipts— Partnerships (billion $/yr) Ratio

2001 1998 66 26 0.394

2002 1999 72 27 0.375

2003 2000 79 29 0.367

The 2003 Statistical Abstract also provides data on gross receipts and net income for “non-farm proprietorships.” Id The ratio of net receipts to gross receipts for proprietorships is 0.433. In Allen , the court used the ratio from partnerships because it was “a figure more favorable to plaintiffs counsel.” 2003 WL 23333580 at *6. Although the party requesting the attorney fee has the burden of proof, defendants were pointed to Allen and given an opportunity to respond but gave no reason why the partnership ratio should not be used. Accordingly, the court again uses the partnership ratio. See Fischer, 214 F.3d at 1121 .

The court notes that the ratio of net receipts to gross receipts for corporations is 0.071. These numbers were not available in the Statistical Abstract the court used in Allen to calculate the average market rate. The court does not use the corporation ratio because it suggests an average market rate of $980/hr, which is clearly incorrect.

The most recent data available from the Bureau of Labor Statistics (“BLS”) describing hourly wages in the San Francisco area are located in “CA PMSA-2002 OES Metropolitan Area Occupational Employment and Wage Estimates” (“2002 BLS Wage Estimates”). See United States Dept of Labor, Bureau of Labor Statistics, San Francisco, CA PMSA — 2002 OES Metropolitan Area Occupational Employment and Wage Estimates, available at http://www.bls.gov/oes/2002/oes — 7360.htm. The BLS provides wage estimates for “Legal Occupations” in the year 2002. See id, “Legal Occupations.” The BLS’ estimates for lawyers are a median hourly wage of $64.88/hr and a mean hourly wage of $70.23/hr. Id. The BLS’ estimates for paralegals and legal assistants are a median hourly wage of $24.64/hr and a mean hourly wage of $25.64/hr. Id. The estimated hourly wages for previous years can be found in Table 2.

As in Allen , the court selects the higher of the median or mean hourly wage because it is more favorable to the party seeking the grant of attorney fees and defendants have not objected to using the higher number. Alternatively, the court could select the lower of the median or mean hourly wage because the party requesting the attorney fee has the burden of proof.

See Fischer v. SJB-P.D., Inc., 214 F.3d 1115, 1121 (9th Cir.2000). But, in this case, the court opts for the former approach.

Table 2. Estimated Hourly Wage Data from the 2000, 2001 and 2002 BLS Wage Estimates.

_Attorneys_Legal assistants_

Mean Median Mean Median hourly wage hourly wage hourly wage hourly wage Data year_($/hr)_($/hr)_($/hr)_($/hr)

1998_41.93_4491_20,77_20,41

1999_58A5_69^5_1945_18.98

2000_5030_5091_2416_23.47

2001_5401_5L33_25.42_24.4

2002 70.23 .6488_2064_24.64

Dividing the most recent mean hourly wage for lawyers, $70.23/hr, by the most recent ratio of net to gross receipts, 0.367, yields an estimate of $191.32/hr as the average market rate for lawyers in the San Francisco area. Similarly, dividing the most recent mean hourly wage for paralegals, $25.64/hr, by the same ratio, 0.367, yields an estimate of $69.85/hr as an estimated average hourly rate for paralegals in the San Francisco area. The court notes that using the lower of the median and mean hourly wage data yields an estimated average market rate of $176.74/hr for attorneys and $67.12/hr for legal assistants.

The court concedes here, as it did in Allen , that the census and BLS data are not likely the “best data available.” See Allen, 2003 WL 23333580 at *6. Nor does the court contend that its simple methodology could not be refined. The court uses this data because the parties have provided nothing better, even after being expressly directed to the court’s decision in Allen .

The court rounds the estimated average hourly rates to $190/hr for attorneys and $70/hr for legal assistants. Rounding is appropriate because the court does not want to suggest that its estimate is precise enough to warrant the use of five significant digits. Nor does rounding unduly diminish the fee award due the fee applicant. The court gave the fee applicant the higher of the mean and median hourly wages estimated by the BLS. For attorneys, choosing the higher number resulted in an hourly rate, before rounding, of $191.32/hr instead of an hourly rate of $176.74/hr. For legal assistants, choosing the higher number resulted in an hourly rate, before rounding, of $69.85/hr instead of $67.12/hr.

The court notes that the BLS data have been criticized, in another case, on the grounds that the data underestimates the average market rate because the data “does not include the self employed owners and partners in unincorporated firms * * See United States Dept of Labor, Bureau of Labor Statistics, Technical Notes, 2002 BLS Wage Estimates, available at http://www.bls.gov/oes/2002/oes— tec.htm. The court finds this criticism unpersuasive because it is unclear whether incorporating partner statistics would cause the average market rate to increase. Although partners at large private law firms likely make more than the average market rate, many solo practitioners likely make less than the average market rate. It is not clear how inclusion of self-employed owners and partners would affect the data.

The data have also been criticized on the basis that they do not include the mark-up required to motivate the owners of for-profit firms. The court finds this criticism unpersuasive because the court, as Blum instructs, determines a reasonable attorney fee without regard to an attorney’s status as a private, non-profit or government attorney.

Accordingly, the court uses $190/hr for lawyers and $70/hr for paralegals as an approximate average market rate.

B

Plaintiff argues that it should be reimbursed at an above-average hourly rate. In arguing for an above-average hourly rate, plaintiff emphasizes the skill and experience of its counsel. For example, in a declaration, Joel Zeldin declares that one of plaintiffs attorneys, Nicholas Waranoff, “has the reputation as one of the top real estate litigation attorneys in San Francisco if not in the entire State of California.” Zeldin Decl (Doc # 52) at ¶ 8. The court does not doubt the skill and experience of plaintiffs counsel but questions the relevance of their skill and experience to the court’s determination of a “reasonable attorney fee” in this case.

As described above, plaintiff is entitled to an above-average hourly rate if it can demonstrate that its counsel were more efficient than reasonably competent counsel would have been. Plaintiff has made no such showing and the number of hours claimed might well suggest otherwise. Accordingly, the court awards plaintiff a reasonable attorney fee at an hourly rate of $190/hr for attorneys and $70/hr for legal assistants.

Y

Plaintiff requests the court to award the cost of internet-based legal research. Plaintiffs claimed costs total $7,596.96. Costs Summary (Doc # 51, Exh 11). For the same reasons outlined above, the court finds this request to be excessive and unnecessary. Because plaintiff provides little detail to its legal research documentation, the court uses the above analysis to determine the reasonability of the legal research cost request. Plaintiff requested 224.7 total hours and the court allowed 136.6 of these hours. Thus, the court allowed 61% of the hours requested. The court deems it appropriate to allow 61% of the costs requested by plaintiff. Accordingly, the court grants plaintiff costs in the amount of $4,618.36.

VI

In sum, the court finds that the number of hours reasonably expended by plaintiffs attorneys is 114.8 hours at a reasonable hourly rate of $190/hr for a total of $21,812.00. Further, the number of hours reasonably expended by the legal assistants is 21.8 hours at a reasonable hourly rate of $70/hr for a total of $1,526.00. Finally, just costs incurred total to $4,618.36. This yields an overall award of costs and fees of $27,956.36.

IT IS SO ORDERED.